Higdon 9803331 The investigator develops improved algorithms for computing numerical simulations of ocean circulation. One motivation for this work is to improve the numerical methods used in the Miami Isopycnic Coordinate Ocean Model (MICOM), but this work also has broader application. MICOM is an operational ocean model that is currently being tested at Los Alamos National Laboratory and elsewhere for use in coupled, ocean-atmosphere models of the global climate system. This model presently uses a three-level timestepping scheme that, in practice, requires a restriction on the time step that is more severe than would be expected from standard theory. In addition, the model is affected by a computational mode consisting of nonphysical grid-scale oscillations, which are particularly stimulated by the behavior of MICOM's representation of the oceanic mixed layer. As an alternative, the present project investigates methods involving two time levels. Such methods have the potential of allowing a substantially longer time step, require less storage, and do not admit a computational mode. These methods are analyzed and implemented in conjunction with a barotropic-baroclinic time splitting, which is used to split the fast and slow motions into separate subsystems that are solved by different techniques. One application of computer models of ocean circulation is to aid in the study of the earth's climate system. For example, a major point of interest in climate modeling is the poleward transport of heat. If the atmosphere and ocean were held stationary, the tropical regions would be hotter than they actually are, and the polar regions would be colder. However, the movement of the atmosphere and ocean serves to moderate these extremes. Roughly half of the poleward transport of heat is due to the atmosphere, and half is due to the ocean. In the ocean, substantial transport is due to bulk movement by currents such as the Gulf Stream; another transport mechanism is the mixing caused by eddies that are shed from ocean currents due to hydrodynamic instability. A proper modeling of these phenomena requires high spatial resolution and computations over long time intervals, and climate-scale simulations of ocean circulation pose great computational challenges that strain the capabilities of the most powerful computers. The aim of the present project is to improve the efficiency and reliability of the algorithms that are used in such computations. A full-scale climate model would require both a model of the atmosphere and a model of the ocean, and these components must communicate in a way that simulates the interaction of the real atmosphere and ocean. The present project focuses on an existing ocean model that is being prepared at several sites for large-scale, coupled ocean-atmosphere simulations.
